Cloud computing has recently emerged as an important service to manage applications efficiently over the Internet. Various 5 cloud providers offer pay per use cloud services that requires Quality of Service (QoS) management to efficiently monitor and measure 6 the delivered services through Internet of Things (IoT) and thus needs to follow Service Level Agreements (SLAs). However, providing 7 dedicated cloud services that ensure user's dynamic QoS requirements by avoiding SLA violations is a

more »

... challenge in cloud 8 computing. As dynamism, heterogeneity and complexity of cloud environment is increasing rapidly, it makes cloud systems insecure 9 and unmanageable. To overcome these problems, cloud systems require self-management of services. Therefore, there is a need to 10 develop a resource management technique that automatically manages QoS requirements of cloud users thus helping the cloud 11 providers in achieving the SLAs and avoiding SLA violations. In this paper, we present SLA-aware autonomic resource management 12 technique called STAR which mainly focuses on reducing SLA violation rate for the efficient delivery of cloud services. The 13 performance of the proposed technique has been evaluated through cloud environment. The experimental results demonstrate that 14 STAR is efficient in reducing SLA violation rate and in optimizing other QoS parameters which effect efficient cloud service delivery. 15 Index Terms-Autonomic cloud, resource provisioning, cloud computing, resource scheduling, quality of service, service level agreement Ç 16 1 INTRODUCTION 17 C LOUDS offer three types of services such as Infrastructure-18 as-a-Service (IaaS), Platform-as-a-Service (PaaS) and Soft-19 ware-as-a-Service (SaaS) and therefore it requires manage-20 ment of Quality of Service (QoS) to efficiently monitor and 21 measure the delivered services to meet Service Level 22 Agreements (SLAs). In Cloud environment, uncertainty and 23 dispersion of resources encounters problems in efficient 24 management of resources, which is caused due to many 25 reasons [1], [2] such as: i) heterogeneity (due to different 26 type of resources and scheduling techniques), ii) dyna-27 mism (detect and fulfill the requirements of application at 28 runtime) and iii) failures (failure of system or resources 29 which leads to performance degradation). However, pres-30 ent cloud computing systems and management techniques 31 are unable to handle above mentioned problems efficiently 32 at runtime. An autonomic system provides a solution to 33 this problem by offering the environment in which appli-34 cations can be managed efficiently by fulfilling QoS 35 requirements of applications without human involvement. 36 Thus, autonomic cloud system becomes self-managed to 37 overcome the above challenges and to provide reliable, 38 secure and cost efficient services to end users. 39 Currently, cloud services are provisioned and scheduled 40 according to resources' availability without ensuring the 41 expected performances [3]. The cloud provider should evolve 42 its ecosystem in order to meet QoS requirements of each 43 cloud component. To realize this, there is a need to consider 44 two important aspects which reflect the complexity intro-45 duced by the cloud management should be considered: first 46 QoS-aware and second autonomic management of cloud 47 services. 48 QoS-aware aspect involves the capacity of a service to be 49 aware of its behavior to ensure the elasticity, high availabil-50 ity, reliability of service, cost, time etc. as mentioned in SLA 51 [4]. Autonomic implies the fact that the service is able to 52 self-manage itself as per its environment needs. Thus, maxi-53 mizing cost-effectiveness and resource utilization for 54 applications while ensuring performance and other QoS 55 guarantees, requires leveraging important and extremely 56 challenging tradeoffs [5]. 57 Based on policy guidance, autonomic system keep the 58 system stable in unpredictable conditions and adapt quickly 59 in new environmental conditions like software, hardware 60 failures etc. Thus, there is a need of SLA-aware autonomic 61 resource management technique which considers all the 62 important QoS parameters like availability, cost, latency, 63 execution time etc. to reduce SLA violation rate for better 64 resource management. In our earlier work [1], [2], [8], [9], 65 we have identified various research issues related to QoS 66 and SLA for autonomic management of cloud resources [1] 67 and based on these issues, we have developed a QoS based 68 resource provisioning technique (Q-aware) to map the 69 resources to the workloads based on user requirements 70 described in the form of SLA [8]. Further, resource schedul-71 ing framework (QRSF) has been proposed, in which provi-72 sioned resources have been scheduled by using different S. Singh and I. Chana is with the I E E E P r o o f 73 resource scheduling policies (cost, time, cost-time and bar-74 gaining based) [9] . In QRSF, manual resource scheduling is 75 considered which further needs lot of human work every 76 time to schedule resources to execute workloads by fulfill-77 ing their QoS requirements. The concept of QRSF has been 78 further extended by proposing energy-aware autonomic 79 resource scheduling technique (SOCCER) [2], in which 80 IBM's autonomic computing concept has been used to 81 schedule the resources automatically by optimizing energy 82 consumption and resource utilization where user can easily 83 interact with the system using available user interface. Our 84 existing research work considers only few QoS parameters 85 of self-optimizing without considering SLA violation rate. 86 This research work is an extension of our previous research 87 work [1] . In this research work, we have presented cloud 88 based SLA-aware autonomic resource management tech-89 nique for both homogenous and heterogeneous cloud work-90 loads and measured the impact of QoS parameters on SLA 91 violation rate. 92